Concepts
Core concepts and architecture of Kage.
Concepts
Understanding these core concepts will help you get the most out of Kage.
In This Section
- Architecture — Supervisor pattern, event sourcing, and system design
- Namespaces — Grouping repositories for context sharing
- Memory — Two-tier memory system for agent context
- Subscription Pooling — Scaling with multiple Claude Code subscriptions
- Approval Workflows — Controlling agent actions with approval gates
- Tasks & Checkpoints — Autonomous task execution with guardrails
- Secrets — Secure credential management via OS keychain
-
Architecture
How Kage is designed and why.
-
Namespaces
Organizing repositories with namespaces.
-
Memory System
How Kage stores and shares context between agents.
-
Subscription Pooling
Scale your development capacity with multiple Claude Code subscriptions.
-
Approval Workflows
Control what agents can do with configurable approval gates.
-
Tasks & Checkpoints
Autonomous task execution with checkpoints, criteria, and guardrails.
-
Secrets Management
Secure credential storage using your OS keychain.